Tesla’s advanced driver-assistance system is now live in mainland China, marking the end of a years-long saga involving regulatory hurdles, data-security concerns, and the kind of bureaucratic patience that would test even the most zen product manager. The rollout began at the end of February 2025 for owners of vehicles equipped with Tesla’s newer Hardware 4 platform.

Here’s the thing: Tesla can’t actually call it “Full Self-Driving” in China. Local regulations prohibit that branding, so the system goes by “Navigation on Autopilot” and “Intelligent Assisted Driving” instead. Same software, different name tag. Think of it as the automotive equivalent of a movie getting a different title for international release.

What the system actually does (and doesn’t do)

Despite the FSD label it carries in North America, Tesla’s system in China is classified as Level 2 autonomy. In English: the car can handle steering, acceleration, and braking in many situations, but the driver needs to stay alert and keep their hands ready to take over at all times. It’s an advanced co-pilot, not an autonomous chauffeur.

That distinction matters because it sets the legal and liability framework. Level 2 means the human behind the wheel is still responsible for everything that happens. Tesla’s system may feel futuristic, but from a regulatory standpoint, it sits in the same category as features offered by dozens of other automakers.

The initial rollout was briefly paused due to heightened regulatory scrutiny before resuming with updated marketing materials that comply with China’s advertising standards. Tesla had to navigate strict data-localization rules that govern how driving data collected on Chinese roads is stored and processed. All of that data must remain within China’s borders, a requirement that forced Tesla to build local data infrastructure rather than relying on its global cloud systems.

That localized data pipeline does come with a silver lining, though. Tesla can now train its AI models on Chinese road conditions, traffic patterns, and driving behaviors specific to the market. Chinese roads are notoriously complex, with dense urban environments, unique traffic norms, and infrastructure that differs significantly from North American highways where FSD was originally developed.

The money angle

Tesla is exploring multiple ways to monetize the system in China. The company is considering a subscription model priced at around $98 per month, alongside a one-time purchase option of roughly 64,000 yuan, which works out to approximately $8,800. That dual pricing approach mirrors what Tesla has tested in other markets, giving buyers the choice between a large upfront payment and a recurring fee.

For context, Tesla has long pitched FSD as a future revenue engine that could generate software-margin income from its existing vehicle fleet. CEO Elon Musk has repeatedly described FSD as the key to unlocking Tesla’s valuation as a technology company rather than a car manufacturer. China, as Tesla’s second-largest market by sales volume, represents one of the biggest potential pools for that recurring revenue.

But pricing the system competitively won’t be simple. Chinese consumers are famously value-conscious when it comes to automotive technology, and local competitors are bundling increasingly sophisticated driver-assistance features into their vehicles at lower price points or even as standard equipment.

A crowded battlefield

Look, Tesla isn’t arriving in China’s autonomous driving race as the clear frontrunner. It’s showing up to a party that’s already well underway. Local players like XPeng and Huawei have been investing heavily in their own systems, and they’ve had the home-court advantage of collecting Chinese road data for years.

XPeng is particularly aggressive. The company has projected that its Level 3 software will be ready within two years, with Level 4 autonomy targeted by 2026. Level 3 is a meaningful jump from Level 2 because it shifts legal responsibility to the vehicle’s system in certain conditions, meaning the driver can genuinely disengage. Level 4 goes further, handling most driving scenarios without human intervention at all.

If XPeng or Huawei reaches Level 3 certification in China before Tesla does, the narrative around Tesla’s technological leadership could take a hit in precisely the market where perception matters most. Chinese EV buyers are increasingly brand-loyal to domestic manufacturers, and advanced autonomy features are becoming a key differentiator in purchase decisions.

Huawei’s approach is worth watching separately. The tech giant doesn’t build cars itself but licenses its ADS (Advanced Driving System) to multiple automakers, spreading its technology across several brands and vehicle segments. That platform strategy means Huawei’s system could accumulate driving data from a much larger and more diverse fleet than any single automaker.

Tesla’s counterargument has always been scale. With millions of vehicles on roads worldwide collecting data, the company argues it has a training-data advantage that compounds over time. The China launch extends that data flywheel into a massive new geography, but the localization requirements mean Chinese data and global data exist in separate silos, limiting some of that theoretical advantage.

What investors should actually watch

The immediate question isn’t whether the technology works. It’s whether Chinese consumers will pay for it. Tesla’s FSD take rate, the percentage of buyers who opt for the system, has been a closely watched metric in North America, and early data in China will be just as telling.

If the subscription model gains traction, it validates Musk’s long-standing thesis about software revenue. If it doesn’t, Tesla may need to lower prices or bundle the feature to compete with domestic rivals who are offering similar capabilities for less.

The regulatory environment adds another layer of uncertainty. China’s government has shown it can pause or restrict Tesla’s operations quickly when political or data-security concerns arise. The brief pause during the initial rollout was a reminder that operating in China means operating at the pleasure of Chinese regulators, a dynamic that introduces risk regardless of how good the technology is.

For the broader EV autonomy race, this launch is less about Tesla catching up and more about the starting gun for a new phase of competition. The companies that crack Level 3 certification in China first will likely capture an outsized share of consumer trust and, eventually, the revenue that comes with genuinely autonomous features. Tesla just put its chips on the table. Now it needs to play the hand.